linux-kernel - Re: PWM regression causing failures with the pwm-atmel driver

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Date:   Wed, 24 May 2023 09:45:27 +0200
From:   Peter Rosin <peda@...ntia.se>
To:     Uwe Kleine-König <u.kleine-koenig@...gutronix.de>
Cc:     LKML <linux-kernel@...r.kernel.org>, linux-pwm@...r.kernel.org,
        Thorsten Leemhuis <regressions@...mhuis.info>,
        Thierry Reding <thierry.reding@...il.com>,
        Conor Dooley <conor.dooley@...rochip.com>,
        Claudiu Beznea <claudiu.beznea@...rochip.com>
Subject: Re: PWM regression causing failures with the pwm-atmel driver

Hi!

2023-05-24 at 08:07, Uwe Kleine-König wrote:
> On Tue, May 23, 2023 at 10:42:34PM +0200, Peter Rosin wrote:
>> 2023-05-23 at 01:34, Peter Rosin wrote:
>>> So, I took a step back and can only conclude that there must be some
>>> another regression to find, and I was confused by that other regression.
>>> In short, I was on 6.1.<foo> and everything was fine, and then I bumped
>>> to 6.3 and a process crashed. I went to 6.2 and that same process also
>>> crashed. I then totally focused on v6.1..v6.2 to figure out the problem.
>>> I simply assumed v6.3 had the same problem because the symptom from
>>> 30.000ft was the same (that process died). I failed to go back to v6.3
>>> to confirm that it was indeed the same problem as I had found in the
>>> v6.1..v6.2 range.
>>>
>>> My bad, it seems I have another day of bisections lined up.
>>
>> For closure, I ended up with this:
>> https://lore.kernel.org/lkml/221d19e2-6b92-7f38-7d8a-a730f54c33ea@axentia.se/
>>
>> I.e. another v6.1..v6.2 regression that caused sound failures.
>> The two problems looked very similar to the suffering application.
>>
>> Anyway, sorry again for the noise.
> 
> OK. After your first mail I had the impression there is another PWM
> releated problem to be reported, but it seems this isn't the case.
> 
> Just to have that explicit: Did I understand you right?

The PWM regression was already solved by 1271a7b98e79 ("pwm: Zero-
initialize the pwm_state passed to driver's .get_state()") and I failed
to notice that. I was simply confused by that other totally unrelated
DMA problem that made me think the PWM problem persisted into the
present when in fact it didn't.

All is well.

Cheers,
Peter

#regzbot fixed-by: 1271a7b98e79